Recent well-documented increases in the rate of new HIV infections in high risk populations coincide with the spreading popularity of club drugs such as MDMA, ketamine, GHB, & methamphetamine. This research seeks to further understand these linkages by systematically exploring the epidemiology of substance abuse among men at high risk for HIV infection. This application will conduct extensive analyses of existing data sets collected on three different probability samples obtained from Chicago, Illinois: The main comparative analyses focus on a sample of men surveyed in 1996-1997 and an additional sample of men surveyed in 2002-2003; additional comparisons will employ a sample of men surveyed as a part of a general population study of drug use in 2001-2002. Taking advantage of parallel instruments and measures across two Chicago-based samples, detailed analyses will compare and explore substance use reporting validity as well as the relative prevalence of substance use/abuse between high-risk and non-high-risk men. Guided by a "cognitive escape" conceputal framework, this application provides a unique, timely, and cost-effective opportunity for understanding shifts in substance use prevalence patterns and trends over time among high- risk men. Multivariate models will systematically examine associations between drug use and risk behavior; associations over time (across different data sets) will be compared'. In order to better understand these linkages, we will explore alternative statistical approaches for combining and scaling survey responses to questions about high risk behavior/including single parameter item response theory (Rasch Scaling) and covariance structure measurement models (using LISREL and MPLUS). These scaling approaches, in turn, will be used to refine and revise statistical models. This application will have a significant impact on public policy by further articulating the causal role of drug abuse in the spread of HIV, thus helping to guide future HIV prevention and intervention efforts. [unreadable] [unreadable] [unreadable]